The synthesis and characterization of two Fe-Gd systems based on bpca(-) (Hbpca = bis(2-pyridilcarbonyl)amine) as bridging ligand is presented, taking the systems as a case study for structure property correlations. Compound 1, [Fe-Ls(II)(mu-bpca)(2)Gd(NO3)(2)(H2O)]NO3 center dot 2CH(3)NO(2), is a zigzag polymer, incorporating the diamagnetic low spin Fe-LS(II) ion. The magnetism of 1 is entirely determined by the weak zero field splitting (ZFS) effect on the Gd(III) ion. Compound 2 is a Fe(III)-Gd(III) dinuclear compound, [Fe-LS(III)(bpca)(mu-bpca)Gd(NO3)(4)]center dot 4CH(3)NO(2)center dot CH3OH, its magnetism being interpreted as due to the antiferromagnetic coupling between the S-Fe = 1/2 and S-Gd = 7/2 spins, interplayed with the local ZFS on the lanthanide center. In both systems, the d-f assembly is determined by the bridging capabilities of the ambidentate bpca(-) ligand, which binds the d ion by a tridentate moiety with nitrogen donors and the f center by the diketonate side. We propose a spin delocalization and polarization mechanism that rationalizes the factors leading to the antiferromagnetic d-f coupling. Although conceived for compound 2, the scheme can be proposed as a general mechanism. The rationalization of the weak ZFS effects on Gd(III) by multiconfiguration and spin-orbit ab initio calculations allowed us to determine the details of the small but still significant anisotropy of Gd(III) ion in the coordination sites of compounds 1 and 2. The outlined methodologies and generalized conclusions shed new light on the field of gadolinium coordination magnetochemistry.